Value Proposition
· Clinical Genomics Gap Addressed: This technology directly targets the systematic underdetection of rare pathogenic germline variants by current industry-standard pipelines, a gap with direct consequences for hereditary cancer diagnosis, trial enrollment, and treatment eligibility.
· Novel Ensemble Approach: Rather than replacing existing variant callers, this technology leverages an ensemble framework that integrates outputs from complementary variant callers to improve sensitivity and precision beyond single-caller approaches.
· Demonstrated Performance: In a metastatic prostate cancer WES cohort, this technology outperformed GATK and DeepVariant in sensitivity and specificity for rare pathogenic germline variant detection. Across known cancer predisposition genes, this technology identified 181 rare pLOF calls that were not detected by one or both individual callers; 29 of these variants were reported in ClinVar as pathogenic or likely pathogenic.
· Disease-Agnostic and Platform-Compatible: The framework is applicable across oncology, rare disease genetics, and pharmacogenomics, and is compatible with existing sequencing infrastructure without requiring re-sequencing.
· Multiple Commercialization Pathways: Deployable as a software module, cloud-based API, or clinical lab add-on, supporting licensing to clinical laboratories, pharmaceutical companies, sequencing platforms, and genomics service providers.
Unmet Need
· Standard germline variant-calling workflows are optimized for broad genome-wide performance, but they can under detect rare clinically actionable variants, particularly predicted loss-of-function variants in cancer predisposition genes such as BRCA1/2, ATM, CHEK2, and PALB2. These missed calls can affect hereditary cancer diagnosis, trial enrollment, treatment eligibility, and downstream manual review burden.
· Rare pathogenic and predicted loss-of-function variants represent a small but clinically high-impact subset of WES data. Because these variants are infrequent and often technically challenging, pipelines that perform well overall may still have reduced sensitivity for the variant class most relevant to clinical interpretation.
· No commercially available tool applies a supervised stacked ensemble framework trained on replicate-validated rare germline variant truth sets specifically optimized for pLOF variant detection.
Technology Description
Researchers at Johns Hopkins have developed a software package which leverages an ensemble framework that integrates outputs from complementary variant callers to improve sensitivity and precision beyond single-caller approaches. The system automatically applies supervised classification to ambiguous single-caller variants improving sensitivity and specificity for rare pathogenic germline variants.
Stage of Development
Fully functional and validated research software pipeline
Data Availability
Available upon request
Manuscript is in preparation
